LASAFT-Net-v2 light¶

LASAFT-Net-v2 Light version for MDX Challenge

!git clone https://github.com/ws-choi/LASAFT-Net-v2 --quiet
%cd LASAFT-Net-v2

/home/wschoi/LASAFT-Net-v2/book/quickstart/LASAFT-Net-v2

# preliminaries: install python3-dev
!pip install hydra-core==1.1.0.rc1 pytorch-lightning>=1.4.1 rich wandb

import soundfile as sf
from IPython.display import display, Audio

mixture, _ = sf.read('data/test/sample/mixture.wav')

display(Audio(mixture.T, rate=44100))

from load_pretrained import get_mdx_light_v2_699

model = get_mdx_light_v2_699()

checkpoint conf/pretrained/v2_light/epoch=669.ckpt is loaded:

result = model.separate_tracks(mixture, ['vocals', 'drums', 'bass', 'other'], overlap_ratio=0.5, batch_size=4)

print('separated vocals:')
display(Audio(result['vocals'].T, rate=44100))

print('separated drums:')
display(Audio(result['drums'].T, rate=44100))

print('separated bass:')
display(Audio(result['bass'].T, rate=44100))

print('separated other:')
display(Audio(result['other'].T, rate=44100))

/home/wschoi/exit/envs/tutorial-environment/lib/python3.9/site-packages/torch/functional.py:471: UserWarning: stft will soon require the return_complex parameter be given for real inputs, and will further require that return_complex=True in a future PyTorch release. (Triggered internally at  ../aten/src/ATen/native/SpectralOps.cpp:664.)
  return _VF.stft(input, n_fft, hop_length, win_length, window,  # type: ignore[attr-defined]

/home/wschoi/exit/envs/tutorial-environment/lib/python3.9/site-packages/torch/functional.py:545: UserWarning: istft will require a complex-valued input tensor in a future PyTorch release. Matching the output from stft with return_complex=True.  (Triggered internally at  ../aten/src/ATen/native/SpectralOps.cpp:817.)
  return _VF.istft(input, n_fft, hop_length, win_length, window, center,  # type: ignore[attr-defined]

separated vocals:

separated drums:

separated bass:

separated other:

%cd ..
%rm -rf LASAFT-Net-v2

/home/wschoi/LASAFT-Net-v2/book/quickstart